String version of col

[maxigit]

col takes Text as an argument which pretty much forces to use OverloadedStrings.
Could it not take a String instead or (people using OverloadedStrings wont notice the difference) or should we introduce a col'` or something which takes a String instead of a Text.

I guess this is for performance reasons, but columns seems to stored internally as Strings so it is a but confusing.

[mchav]

I had a separate lib that used String everywhere and got this comment:

String being the default type available from base is a historical accident that other languages have had (like Erlang), and the availability of text in GHC distributions should not be a blocker to adopt it. It does not make sense to use String in 2025.

I like string better tbh but suporting both would be tricky. col' would be a good function to add but the problem would be that most of the places you use columns you end up also using text e.g. in derive. Also for all the stuff stored in the dataframes string would make it slower than it already is (it's already kind of slow).

There is a microhs prototype which uses string and I will say it's nice not having to worry about turning on overloaded strings all the time.

That one also incidentally exports everything using one namespace.

[maxigit]

Fair enough, I though that I've seen that DataFrame was storing the column as a Map of String but I am wrong.

String being the default type available from base is a historical accident that other languages have had (like Erlang), and the availability of text in GHC distributions should not be a blocker to adopt it. It does not make sense to use String in 2025.

This is one person opinion. String has many advantages over Text and is perfectly fine for things which are not critical. Using String for the legend on a plot seems reasonable.